In this paper, an accurate and effective method is proposed to construct the driving cycle of electric vehicles in urban area. Taking Xi'an as the research object, the driving cycle of electric vehicles is constructed. The method mainly consists of four parts: developing test routes, data collection, data processing and driving cycle construction. Firstly, according to the overall road topology and traffic flow survey results of Xi'an, the test route is developed. Then, collecting the driving data. Aiming at the problem of complex process and low clustering accuracy in traditional fragment partitioning, dimension reduction and clustering data processing methods, the velocity-acceleration grid is proposed to divide the original data directly. Moreover, in view of the single classification condition and the loss of small probability events in the traditional Markov method, the Monte Carlo simulation optimization method is proposed. Finally, the construction of the driving cycle of passenger cars in Xi'an is completed. Compared with the driving cycles built by other methods, the driving cycle established by using the proposed method is relatively accurate.

CICTP